Re: <xoi> and fuzzy models

[Jorge Llambias <jorge@PHYAST.PITT.EDU>]

la stivn cusku di'e

> A challenge to all: produce a gismu which can not be considered to be fuzzy.

There are some (e.g. dugri, tenfa, sinso, tanjo) for which it wouldn't be
very easy to consider them fuzzy.

> le rozgu pafi'uci xoi barda le ka melbi kei le mi'o ckilu
>
> The rose is fuzzily 1/3 big in property beauty on our (predefined) scale.

Rather:

The rose is big in property beauty compared to our scale.
(truth value of that sentence = 1/3)

The x3 of {barda} is not for a scale, but for something with which to
compare the x1. You would be saying that the rose is big in beauty compared
with how big our scale is in beauty. > If none had them we could deal with
 standards using some other formalism.
> Among other benefits, such as regularizing all gismu, this would allow us
> some more options in expressing fuzzy sets.

Right. And what's more, the mechanism for that is already there.


> If the color was part-way between two colors, could we define the sumti
> place structure to have a scale sumti, say at X2?
>
> le rozgu refi'umu xoi xunrypelxu le mi'o ckilu

I don't think so, but you can always use {ci'u}:

        le rozgu cu xunrypelxu ci'u le mi'o ckilu
        The rose is red-yellow on our scale.

> The rose is 2/5 fuzzily red-yellow on our scale (that is, a more reddish
> shade of orange, in the American English color idiom).

Then what you want is probably:

le rozgu cu xunrypelxu sela'u li refi'umu ci'u le mi'o ckilu
The rose is red-yellow in quantity 2/5 on our scale.

I don't think you are talking about fractional truth values here.

> leti rozgu pafi'uci xoi melbi leta rozgu fo le mi'o ckilu
> This rose is fuzzily 1/3 more beautiful than that rose on our scale.
>
> leta rozgu pafi'uci xoi melbi letu tujli le ka sumne kei le mi'o ckilu
> that rose is fuzzily 1/3 more pleasant than yonder tulip in odor on our scale.

But the place structure of {melbi} is not "x1 is more beautiful than x2".
The x2 of melbi is the person for whom the x1 is beautiful. It seems to
me that you are not using {xoi} for truth values, either, but rather for
values on a scale. For that there is {sela'u}.

> Either every gismu should have both a standard sumti and a scale sumti or
> no gismu should have them. The current seemingly random position and
> existence of standard sumti makes no sense.

I agree completely. Notice that you can always add a standard with {ma'i}
and a scale with {ci'u}.

> If all gismu had them, then one
> could say stuff like:
>
> X1 <selbri> X2...X(N-2) by standard X(N-1) on scale X(N)

You can already say that:

X1 <selbri> X2 X3... ma'i <sumti> ci'u <sumti>

Jorge